An apparatus and method for facilitating a subsequent choice of a motion vector from a plurality of candidate motion vectors in a motion estimation method, include receiving video image data from first, second and third consecutive groups (10, 20, 30) of image data. A first image part (12) of the first group of image data (10), a second image part (22) of the second group of image data (20), and a third image part (32) of the third group of image data (30) are identified, wherein the positions of the three image parts (12, 22, 32) correspond to a motion trajectory that is indicated by a candidate motion vector. A first error measure is then calculated by testing for differences between the first image part (12) and the third image part (32) and quantified as a first error measure. A second error measure is calculated by testing for differences between either the first image part (12) and the second image part (22) or the second image part and the third image part and quantified as a second error measure. An output error measure is then quantified, wherein the output error measure is derived from either the first or the second or a combination of both error measures.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A method for facilitating a choice of a motion vector from a plurality of candidate motion vectors in a motion estimation method, the method comprising the steps: receiving video image data from first, second and third consecutive groups of image data; identifying a first image part of the first group of image data, a second image part of the second group of image data, and a third image part of the third group of image data, wherein the positions of the three image parts correspond to a motion trajectory indicated by a candidate motion vector; calculating a first error measure by testing for differences between the first image part and the third image part, and quantifying the differences as the first error measure; calculating a second error measure by testing for differences between either the first image part and the second image part, or the second image part and the third image part, and quantifying the differences as the second error measure; and outputting an output error measure, wherein the output error measure is derived from the first error measure and/or the second error measure in dependence on a vector length of the candidate motion vector.
2. The method as claimed in claim 1 , wherein when a candidate motion vector length is above a maximum length Vmax, then the second error measure is selected as the output error measure.
3. The method as claimed in claim 2 , wherein, for candidate motion vectors having a vector length greater than Vmax, the third image part is not used, whereby the output error measure only comprises the second error measure and the corresponding first error measure need not be calculated.
4. The method as claimed in claim 1 , wherein where the vector length of the candidate motion vector is between zero and a maximum length Vmax, the output error measure is a value based on a combination of the first error measure and the second error measure.
5. The method as claimed in claim 4 , wherein the output error measure is given by the equation Error out =f(L)*Error pc +(1−f(L))*Error pn , wherein Error out is the output error measure; Error pc is the second error measure; Error pn is the first error measure; and f(L) is a function of the vector length of the candidate motion vector, f(L) being one when the vector length is greater than or equal to Vmax, and being zero when the vector length is zero.
6. The method as claimed in claim 1 , wherein the output error measure is calculated by taking a weighted sum of the first and the second error measures.
7. An apparatus for facilitating a choice of a motion vector from a plurality of candidate motion vectors in a motion estimation device, the apparatus comprising: storage means for receiving and storing video image data from first, second and third consecutive groups of image data; and processing means for identifying a first image part of the first group of image data, a second image part of the second group of image data, and a third image part of the third group of image data, wherein the positions of the three image parts correspond to a motion trajectory indicated by a candidate motion vector; said processing means calculating a first error measure by testing for differences between the first image part and the third image part, and quantifying the differences as the first error measure; calculating a second error measure by testing for differences between either the first image part and the second image part, or the second image part and the third image part, and quantifying the differences as the second error measure; and outputting an output error measure, wherein the output error measure is derived from the first error measure and/or the second error measure in dependence on a vector length of the candidate motion vector.
8. A video processing apparatus incorporating video motion estimation utilizing a motion estimation method in which candidate motion vector selection is facilitated by the method as claimed in claim 1 .
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
February 19, 2002
April 4, 2006
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.